things, any democratic consultants i talk to when i'm running for office will say. it is kind of like okay, we spent a lot of time convincing ourselves we are right about everything we believe. what i suggest is why don't we bring in a republican strategist? not everyone would do it but we find someone who would do it, bring in someone from a think tank or opinion writer from the washington post with a more conservative bent, listen to how they are thinking about the world and politics. maybe we will learn something. >> afterwords airs on booktv every saturday at 10:00 pm eastern and sunday at 9:00 pm eastern and pacific. all previous afterwords programs are available to watch on her website booktv.org. >> thank you for inviting me. i study airplane crashes because they are interesting and educational. i'm not here to suggest flying is unsafe and this is how i

normally introduce myself quickly, so i wrote two books about airplane crashes, the first was turned into a training seminar, you will over the lights please. second one just came out, co-authored with a real airline pilot and my presentation is based on the most dramatic videos i could find, large planes get struck by lightning once a year and all planes will be destroyed by a thunderstorm but it is a small plane problem. it hasn't happened in the modern jet era. small planes with an experienced pilots without weather radar will wander into a thunderstorm. it is a lot of simple things such as no loose rivets and vapor spaces in the fuel tank

allowing arcing. so it has a big plane in the united states hasn't been destroyed by lightning since 1963. i like to start with a lot of dramatic pictures, the wake turbulence, mini tornados many that behind a boeing 777, could be 200 miles an hour and found a couple examples of small planes getting too close to a big plane, being broken up in flight and only take off so close to each other, most definitely limits throughput. in 1972 a dc-9 was flipped over and crashed 21/4 miles behind a much larger dc-10 and increased the separation space to 5 miles and i used to do a lot in this talk with explosive decompression including two people who got sucked out of

windows, so keep your seatbelt on, you are supposed to do that anyhow. a few people have been known to break their neck and bouncing their head off the ceiling in severe turbulence. a plane is fine, the people can rattle around. this 14' x 18' blast whole, beyond any rational design. it also says something, quite a few planes survived in flight bombings and managed to safely land which everyone had their seatbelt on and lived except one unfortunate flight attendant standing near the blast whole and the problem is metal fatigue, every big plane you cycle enough it will fall out of the sky. they know how to prevent it. things still happen for a variety of errors mostly manufacturing defects so to

scale ribbing on at 737, a highly redundant structure that had something to do with the plane surviving, remarkable it didn't break in two. one of the first things i like to explain is how survivable that airplane crashes. a survey of all dc-10s made, 27 destroyed in crashes, four had a total loss of life, terrorist bomb, two mechanical failures which are engineering defects of sorts. 85% of the time, 90% of passengers survived even if the plane is broken in two or three places and that is because you can always crash hard enough to kill everyone but simply not the most common scenario. they typically crash during botched takeoff or landing at much lower speeds making them far more survivable.

here is a made-for-tv crash. 727, believe it or not this is a highly survivable event, you can get whipsawed violently at the fuse was break and if the seat is released it is like a high-speed car crash without a seatbelt on but i am saying this crash is more violent because it has two fusil is breaks and recently good outcomes. which brings me to this miracle crash of a large built-in 11 in the everglades, usually when you go to the crash site you can say there is the fuselage, might be a few pieces but in this case there was a scattered debris field, no recognizable

cross-section and the crash was judged non-survivable. if judged survivable it triggers a different engineering study, how well did the structure do to protect the people? only problem here, 77 people survived, nobody knows or cares that there is no attempt to make any sense out of it, nothing could be learned from a plane that is badly fragmented. this is perhaps the most severe accidents i'm aware of, this dc-10 had an uncontained engine failure in the engines section and lost all 3 hydraulics embossed control of the flight surfaces and one of the few places where all three hydraulics separated on purpose

for this very reason. a very severe crash, 75% of people survive impact. and subsequent fire as instructed which is the fbi studied it and they just came out last few years within infant seat and the faa concluded if they made parents by next her ticket they would be more likely to not fly and kill more people in highway accident so standard procedures place a baby on the floor so there is the ground scarf showing how severe the impact is and we do have a video. they scrambled, since all the nurses and doctors to the

hospital and they do not land in sioux falls but coincidentally they had a drill in the last year so they were reasonably well rehearsed. the most severe survivable accident i am aware of. here is the actual titanium hub designed for metal fatigue, 18,000 cycles thrown away and replace it, it is good for considerably more cycles and a 1-time overspeed from the centrifugal load and developed no probabilistic design methods based on fracture mechanics and faa mandated metric is one failure per billion flights and was a casting defect, on the

right and afterwords they studied in the simulator could human controllers flight so no control surfaces. all they do is alter engine thrust and concluded no. nasa developed software to steer the plane with altering engine thrust but the faa never mandated it. the faa had all other federal regulators evaluate new safety procedures and equipment which went to cost-benefit analysis and do it in secret involving putting price on a human life, something no company can do because they would get rained in court and the court of public opinion so if the problem almost never happens the cost-benefit analysis is going to conclude there is little benefit and you can justify spending almost $0 and

there are hundreds of things that almost never happened. in my opinion it makes it difficult to get rid of the pilot. more cost-effective to keep the human in the loop for many many things that almost never happened. i studied a more recent uncontained failure. they had massive damage, massive random damage, the computers were the heroes but so were the human pilots. if you try the design for all these things that almost never happen you end up with million-dollar tickets. the engine manufacturers very competitive and fortunate enough to have a test video or

audio. [person making noises] >> that is enough of that. a ge engineer sent that to me. so let's compare a more modern digital boeing 777 and analog dc-10. every generation airplane has more computerization, more computer oversight, they connect more sensors to the computer and safety has improved. a modern cockpit is a bunch of computer screens that holds analog dc-10. here is some safety statistics on the two, a similar number of

plane mirrors and somewhat surprisingly the boeing 777 has less fatalities but not as much as you might expect or hope for. it turns out there were two events with total loss of life, one was a missile strike in the ukraine, the other the missing malaysian plane so we will dismiss both of those is criminal acts which are risks of flying but we are engineers, it is not our fault. if we subtract criminal acts there's a dramatic difference in safety between the two, only three died in a boeing 777 x in two of those three determined they would have lived if they had their seatbelts on and the third one was ambiguous on that

point. so you definitely have increased safety with more computers and the computers never caused a crash, however, but, but, but we can probably say the computers have added to confusion which we will talk about a little bit. here is the most serious 777 crash, a second plane destroyed by a crash except for criminal acts and the pilots lost track of software settings, they thought.a throttle, the plane lost speed, crashed and broken the fire but the structure did a good job of protecting the people, only 3 died and only two died because they didn't have their seatbelt on. it is possible a third one was a seatbelt issue too so this was a very unusual flight, turns out to be a student

pilot, inexperienced airline pilots, first time landing a boeing 777, a very experienced pilot, first flight as -- instrument landing system shut off for maintenance which meant the computer would attract the glide path, there had to be old solutions which had been around decades, there are rules of thumb for a stabilized approach in terms of speeds and altitudes. it is flying 101 and another low-tech system they can use that has been around forever. if they are on the proper glide path they get two white lights and two red light and if they drift up or down the number of white red lights changes so that is what they are supposed

to fly but it was more of a manual flight than they usually do. here is another example, the computer confused things, this airbus in 2001 took off from canada about for portugal, runs out of fuel in the middle of the atlantic and ends up having an entire glide for 75 miles on an abandoned military runway in the azores so they had absolutely no power, there is a solution for that condition, emergency winter pops out and supplies power for controls and the first indication of something odd, there is a fuel week, flow by the berry oil, heat exchanger with the fuel,

that chilled -- 149 ° and that confused the pilot so the pilot died in indicator which is not a warning or an emergency but this is going on here so they didn't know what to make of that and one of the things they can do is call for help, engineers on standby, to help them out, they were confused too, they were stuck on computer glitch, they assumed it was a computer glitch even though they get indications with the fuel system and they ignored them until the first and inflamed out and had to admit they were running out of fuel and the second engine flamed out. so my offer copilot airline

pilot who was a little older and said before everything was computerized they used to teach pilots if you got a little oil temperature think leak. we have some common enough, engineers lost track of it. in my mind that is a bit of an engineering screwup. airline pilots teaching look through the computer screens, everything is computers. another variation with a far worse outcome is now a leading cause of disaster for big planes. of the autopilot shut down for a variety of reasons, we talk about a specific example and you combine that with confused pilots and both are very rare, neither one is supposed to happen. if you have both of them happen you can end up with loss of

control and in my mind this is playing out once they start the computer glitch it becomes confusion magnifier. who hasn't had weird computer problems nobody understands? here is an example of this. a big airbus leaves brazil bound for paris, air france flight 447. torpedo tubes freeze, shuts off the autopilot with an alarm, mr. pilot, take over. this is remarkable. how long have they been fine airplanes at 40,000 feet and freezing conditions? it turns out snow and ice are highly variable and not repeatable and you have a complex distribution of droplet

sizes between missed and rain, freezing on contact and temperature altitudes wept, there's a lot of variables in play. if they did testing in a wind tunnel, who is to say they had the worst combination of circumstances and if the autopilot shuts off combined with confused pilots, this is what remains of air france 447 after slamming into the ocean at 180 ft./s, this is one possible scenario for the missing plane and there's nothing to be found on the surface, 3% of the plane, bits and pieces floating on the surface so all this happened just for the half minutes after the autopilot shuts off, they

were crashing into the ocean. i will begin this story in 1932. the army air corps did a big study and concluded experienced pilots could not fly straight and level if they couldn't see the horizon so they invented the artificial horizon and air france flight 447 was a dark and stormy night, they cannot see the horizon. everything was working except the autopilot and airspeed indicator so they had enough information to solve the problem, they just had to maintain straight and level flight and could not do it which in my mind is the same as what they concluded in 1932. you turn in a car there is a sideways shove. a coordinated turn in an airplane is a little different. the inertial vector of the circular motion adds to the

weight vector and it goes straight through your spine, no sideways tilts. if you have a glass of water on a tray in front of you and your banking the plane 30 °, you cannot tell by looking at the glass of water. that spatial disorientation is another complicated thing but feeling a turn does not help and adds to the confusion and they had intermittent stall alarm going on and off, validator speeds, torpedo tubes freezing and unfreezing and refreezing and they were confused and couldn't figure it out, they should have figured it out. the pilots over steered the plane. if your steering goes bad on your car what do you do?

oversteer trying to get a response? so they went from a very aggressive climb to a dive and stall, turns out when you are in a dive, if you are stalling you have to lower the nose putting in a deeper live to lower the angle of attack and get out of the stall which defies common sense. that is not what humans want to do but experienced pilots are supposed to know better and get enough information from the artificial horizon that solves the problem. the cockpit voice recorder recorded turbulent buffeting indicating a stall and most of the 41/2 minutes once they headed into a dive their angle of attack was 33 ° or higher which is a bad stall in the pilots never discussed stall so

they could hear a recording of the turbulent buffeting and vibration. the turbulent air coming off the wings will shake the tail section, make noise and somewhat surprisingly boeing has a stick shaker, part of the feel of flying, the shaking of turbulent air from stalls, shaking the tail, comes through to the control yoke and shakes and bowing artificially creates that to manage when the pilots take a stall. and airbus does not have a stick shaker. and prevent stalling, a boeing

autopilot does not. they had a degraded autopilot, the computer cannot fly the plane without airspeed, torpedo tubes measure airspeed or frozen and unfreezing. it turns out pilots, big plane pilots never practice stalling the plane. they teach small plane pilots, it is considered unsafe to stall a big plane so the simulator flies up to impending stall when the stall alarm goes off. because of this accident and loss of control the faa mandate reengineering all the simulators by 2019, they can practice it but a massive engineering job, a few test pilots stall the 737/700 times and nasa did a big study, to

replace the test flights with wind tunnel tests and concluded yes, but still needed windtunnel which is indicating limitations of cfd, not my area of expertise. he would think the world would want to recover this accident and there has never been any attempt whatsoever to study the loss of control so i assume it is beyond the capability but they have to reengineer every plane model, it is a very big engineering job and the faa mandates they need to have the new simulator certified with test pilots and there are no test pilots with experienced stalling planes. my co-author airline pilot is an airline instructor and him and i both have heard nothing in two years so assume they won't meet the deadline and the europeans are not talking about

this at all, it is a massive engineering problem, won't be solved easily. the official conclusion, official reports done by the french of air france 447, no explanation why three experienced pilots could not maintain level flight. my theory again, pilots cannot maintain level flight unless they can see the horizon, they see the artificial horizon working and my theory once they start thinking computer anomaly whatever confusion they have is magnified but that is just me talking, not what the official crash investigation says. ice has not historically been a problem in other places, the engines and wings, they did a

lot of new fundamental research and the faa put out new ice rules but who is to say they have the worst combination of circumstances now. want to say a little bit about the missing plane. one theory is hypoxia and there is a precedent for that, boeing 737 crashed, everyone passed out from hypoxia, that is the plane being buzzed by the greek air force, everyone passed out on board and there is one switch that turns on the ventilation system for the old analog 737. the switch is always on auto. the only reason to put it on manual is mechanics put it on manual for maintenance which they did the night before and faulted boeing for not calling out the switch more specifically on the mechanics

just list, checklist. they told the mechanics to restore the control panel to its original configuration, did not call out the switch, alternate is for a second controller and also looking at the sun and low-pressure alarm went off at the same sound as you forgot to retract your flaps shortly after takeoff alarm, so the greek pilot in german copilot by debugging the wrong problem and not native english, while going hypoxic and they passed out, the plane for one autopilot until it crashed so i talked to two 737 pilots about this accident and seem to know more about this switch than they did. they never touched the switch and what is going on here? did i find the world's two dumbest 737 pilots? i am thinking, they don't let

stupid people fly these planes. reminds me of the time i spent a week on an aircraft carrier. there is me steering of aircraft carrier so everywhere i look my head was spinning with engineering questions was my host said to me stop bugging us with engineering questions which our job is to follow the procedure. engineers write procedures, mechanics and pilots follow-up and sometimes bad things happen. the missing plane changed direction in a major way twice and also had numerous minor course corrections, someone was a pair joyriding and also a lot of electronic equipment stopped transmitting, fundamentally there's radio so no reason ever to lose contact and transponder

transmitting, atx hash b is a new gps upgrade on the transponder. they'll season pedestrian to mechanics to get ready to fix this because they can't make any money without the parts being repaired. so it was pinging the networks once an hour saying keep listening to me, keep listening for me. so with the air france 447, they halls a fall transmission every ten minutes and that narrowed the search area down bigger than

the size of connecticut, and the official search area nor missing plane us bit the size of ohio. trying to make sense out of the state's compared to australia a, >> superimposed on the united states and the indian ocean so the official search stopped after they searched the state of ohio. a apparently a private company restarted. i don't know why they think they know anything. that person tried to disappear and if i tried to disappear, i'd change course a few additional times. so i can't see how they are ever going to find anything. and i dade lot of fun with crash testing. it turns out the human body and the plane respond differently to vertical versus horizontal crashing. very little tolerance for

vertical sink rate, hopefully you'll have a long flight out to dissipate your horizontal kineticking. they're kind of crashing down more than they're designed to do. only been three full-scale crash tests in history. there's been a few more fuselage drop tests. it's not mon -- man date but there's more of those. this is a section of a fuselage dropped 6.2-foot, impact speed of 20 feet per second but that's a botched takeoff or landing, this is the biggest, most recent fullscale crash test and there's a video of that. this is a rather -- so, this --

the impact because just 17 feet per second so that's highly survivable. they are doing -- testing an additive to fuel to prevent it from misting and prevent post crash fire. it failed miserably. they wanted to rip through the wing tanks but instead they slid through -- they ripped open he engines and exposed the hot engine parts to fuel dumping on it. now the fuel is not inflammable and had a long chain molecule added to prevent misting and somehow the faa estimate 19 of the 53 simulated passengers would survive the fireball. boeing did a survey and pilots report a hard landing at just four feet per second. that means mr. mechanic, we hit

a little harder than normal. take a look at things. the design limit on the landing gear is 10 feet per second which is only a 19-inch don. you expect damage to the plane, and you're expected fatalities above 25 feet per second. this is based on a survey from the early '8's and thinks have got an little safer, i think, and how safe? that kind of information interact width the lawsuits so you won't find hard numbers them harder you hit most likely to have a post crash fire so there's two event, the impact and post crash fire. and that miracle crash that badly fragmented plane was 37 feet per second, and that's a stat on 100% fatalities.

so here's an example of i think things slowly betting better. the 787 is compass -- composite and the fa did not mandate a crash test, and there are crash-worthy rules. proffer it's as crash worthy yes a aaluminum and boeing dade 30-feet-period-second drop test and the test dummies pasted. that's a lot better than expect ing fatalities at 25 feet per second, turns another ejecting up is the same as crashing down. both come press -- compress the spine and this ejection from an air force thunderbird buy lot during an hair shore. barely ejected ahead of the

crash. nobody got hurt. there's an altimeter here. a thousand feet closer to the ground than supposed to be so it's a show-opening dive, and it's faster than he can make sense out of the instruments. he send something was wrong and ejected. and he some shrunk by two inches and the air force did a study and concluded 7% of ejections end in spine fractures. there's a variety of scenarios where you eject from the plane. the problem is ejection seat ejecting up provides acceleration and if you you're pulling out of a dive. , that's center, seeking acceleration that adds to the ejection seat acceleration so the best they can manage that situation is to result in 7% spine fractures of ejections. kind of have no control over what the plane is doing pulling

out of a dive. the germans studied this during world war ii and started making high performance fighter planes, and they concluded 25 gs where crush your spine which converts to 1500 pounds which is a metric still used by the faa today. they have been mandating crash testing of seats for i think maybe 10-15 years, which is obviously a lot easier than crash-testing an entire plane. so the passed the seat crash test. the crash test dummy has a load global he spine and can't exceed 1500 pounds, and believe it or not, the human is not the weak link. the spine is good for 25 gs, but the only crash test the seat to 14 g vertical impact, and this next video shows why it

doesn't matter. basically -- well, this is a cargo plane flying for the military in afghanistan, and basically if the buyer plane shatters around you, doesn't matter how strong the seat is. and if you hit hard enough to spray fuel, you're going to ignite a fireball, and there's a shadow of a boeing 747. so, they were hauling around 5 heavily armed, heavy military armored vehicles, and they weren't properly latch lashed, and they shifted, there's very strict limits on the cg limits of any airplane. for a boeing 747 it's typically

has to be within the 4-1/2 feet of that yellow zone, and because of the shift, it shifted six feet aft of the rear limit, and somewhat surprisingly, boeing concluded the plane was in fact recoverable from such a severe shift. you needed additional damage to the elevator hydraulics so basically there's limited control of the flight so the severe cg shift, plus damage to the elevator controls, crashed the plane. somewhat interestingly, if you had a cg off that much, a plane would never take off. it would rotate very, very early, and the pilots would abort the takeoff. but apparently in flight it's recoverable if the elevators are working as designed. here's the first boeing 777

crash, and shows what a 25-feet per second sink rate looks like if you have your landing gear extended. turns out an approach is mostly an unpowerred glide. the engines are at idle. but the last minute, once you lower the landing gear and the flaps, the drag increases significantly. so they need thrust at the very end and the auto throttle lost thrust attitude. so doesn't have enough speed to maintain safe flight. so, boeing estimated the stall at 140 not but don't exactly know, and it was down 180 knots hassing high of 200 feet, and basically they don't totally understand stalling close to the ground.

there's aground -- a ground effect so they'll study accident width the simulator, and the simulators suffered with test flight data, and it's not safe to stall a big plane close to the ground. so, they can't exactly simulate stalling close to theground. they can estimate it. and believe it or not, the problem was fuel line freezeup. this is a shell and tube exchanger between the fuel and the engine bearing lubrication, oil. they did data mining of 170,000 flights, which is first time they studied an accident that way so this is a new approach, and they concluded it was the worst combination of three things. it wasn't the coldest flight, it was among the two percent coldest and minus 101 degrees f flying over siberia, and it had

very low thrust during cruise. a nighttime flight, trying to avoid thrust to not wake up the sleeping passengers and that's important because high thrust -- there's ambient ice dissolved in the fuel it and will form on the fuel system piping, and when you hit the thrust, they'll knock it off. so if you don't have a lot of thrust during a cruise, you'll accumulate more ice, and they had a particularly high thrust on approach. so they decided it was the worst combination of those three things, or 175,000 flights and did redesign the shell and tube exchanger. that's what impact 259 feet per second looks like with the landing gear extended. shoved the landing gear through the wings. one person had a bren leg -- a broken leg.

that was the only injury. i look at dutch roll as it is historically interesting and there's a few recent examples. the first large swept wing plane was a bomber designed for boeing for the air force, and it wouldn't fly straight. so boeing solved the problem with the yard damper which was controversial at the time because most engineers believed that planes ought to be stable on their own, shouldn't need a mechanical device that could break. they've since designed dutch roll out of planes. they are in fact stable on their on but took a few -- a couple of decades to do that. so, i'm not sure this animation shows enough yawing-but you'll have a little bit of yawing triggered be -- by the random gust and gate lift unbalance.

the forward wing will have more lift than the trailing wing, which will induce roll so the boeing 707 had pretty severe dutch roll initially, and they were doing dutch roll training recovery in 1949, and they rolled so badly they threw three of the four engines from centrifugal loads and had a bad crash. there are few other similar accidents, and this is back when they were killing pilots training see anywhere yore that never actually occurred with during flyings with passengers and this is before the simulators got as good as they are. so, they made some modifications , tex johnson, famous boeing test pilot who is famous for rolling a 707-barrel roll upside-down in front of a

few hundred thousand people in seattle. so, they increased the size of the vertical stabilizer and added eventual fin. they'd did other things butted the idea was it would increase the resistance to yawing. so, if a bigger tail can suppress dutch roll, perhaps a smaller tail can induce dutch roll, and that happened in a modern airbus at 2005. the composite rudder fill off because of manufacturing defect, and the plane immediately started dutch rolling. the pilots were able to recover and it landed safely, and the plane got more stable in lower, denser air, unlike this more recent flight of air force tanker, which is a boeing 707 -- a modified boeing 707. the yaw damper broke, and the

plane entered dutch roll, and they decided the pilot wasn't properly trained to recover from it. you're not supposed to use the rudder. the problem if the plane -- well, a plane is not designed to fly sideways. if it yaws too much and you swing the rudder in the opposite direction it scoops out more air than it's designed to do and you'll do structural damage to the tail which happened to the ck-135 and the plane crashed. and this airbus crashed a month after 9/11. it wasn't dutch roll but it was flying too much sideways. yawed to much and broke off its vertical stabilizer and crashed. so it was a case of excess yaw and full rudder break offering

the vertical stabilizer. this is also the first and only composite structural failure that crashed, big commercial jet. so, vertical stabilizer was attached with six lugs and they broke first. so this one has 150 plies or carbon come -- composite. the carbon fiber is small, 15 to 25 times finer than a human hair depending on whether they're swedish or italian. so, this plane was flying behind a 747, and it was being postled by the weight turbulence. the weight turbulence bid not cause a structural problem about the problem was the pilot was swinging the rudder back and forth as fast as he could for passenger comfort because of the

postling of the weigh -- the postling of the eight turbulence, and the plane's design for full rudder right or full rudder left, not designed for the rapid cycling of the rudder back anding for as fast as you can do. so, what happened when you have full rudder right or left the plane overshoots a little bit before it oscillates to he equilibrium position and the overshoot was too great a yaw, flying too mach sideways, and when they swung the rudder in the opposite direction rapidly as fast as he could, combination of too much sideway flight with full ruddser in he wrong direction, excess structural loads on the tail, broke off the vertical stabilizer. so, this was a series of improbable events. the weight turbulence is a little odd.

not that unusual. but engineering screwup and the engineers wrote the flying manuals and did not specifically say, don't do that. don't swing the rudder back and forth as fast as you can. there's a with of a disconnect, too. the engineer said we wrote in the flying manuals for the way you're supposed to fly the plane. you're not supposed to do that. and the pilots thought, well, gee, there's nothing we can do to break the plane structurally, right? and every a flying manual called out, don't do that. just this model failed to do that and that's an engineering screwup. there's another situation unique to this plane. pilots controlled -- complained but the pedal forces, so on just this model of airbus, they altered the pedal forces, made it considerably smaller force,

so just on this model, it was easier by a lot to swing the rudder back and forth as fast as you can, and it was considered aggressively -- aggressive piloting. they called it's cowboy maneuver, not supposed to swing the rudder back and forth as fast as you can, but the flying manual didn't call it out. now it's emphasized in training but took 15 years for that combination of circumstances to line up and break the plane. more or less they solved all the ordinary things. you practically have to invent a new combination of circumstances to crash a big jet. there's still pilot air you're but the engineer things are mostly solved. you have to create a unique set of circumstances.

may all your decompressions be slow, an old irish toast. any questions? [applause] i think we are supposed to ask you to go to the microphone to ask a question. suppose i can repeat the question. any questions? >> i wanted to ask about on that airbus that was going over the atlantic, i think -- >> yes. >> what happened in terms of why did it run out of fuel, like, other than -- >> well, they replaced the engines at 2:00 in the morning. most bad things happen at 2:00 in the morning without proper documentation. the engine had been in storage for a year and a half, and unbe

unbe noted to them, they mixed and matched old and new likes and when they installed okay and when the pressured, it was a very big fuel line leak. >> very interesting talk. was wondering -- maybe too soon to know anything but the southwest flight that happened recently where the woman died because the part of the engine hit the window, i yes. >> do you've have any thoughts or know anything about the state of that study? >> well, i typically ignore events until they come out with a final report, and early reporting is usually way off the -- i did have two examples of people being sucked out of

windows. i usually do in this presentation. i took them out because of the currency. so as far as i know it's happened three times. it's extremely rare event. if you have stuff flying out of the engine, the most likely outcome is nothing. it's flying off on a 360-degree arc, so most of the arc is empty stays so you have to be realun fortunate and i do know she wouldn't have been partially sucked out of the window if she had her seatbelt on. other than that it don't know anything. seems to be a lot of uncontained failures recently. it's not supposed to happen. but it does. of course, there's more flights as time goes by. >> i was wondering, about the

malaysia 370 flight, how you said it's -- you think that somebody intentionally made that disappear and made got down. how do you rule out a mechanical failure, even like a fire the cockpit? how could you assume that one one of the issues. >> well, first of all, i think the entire aviation world understands it is a criminal act and the official investigators will not say anything without hard facts. so here they are. think everyone knows what happened and they won't say anything to adds to the mystery. so two things that can explain it, hypoxia. i beat that one to death. fire could knock out equipment and make it stop transmitting, but if you're on fire -- well, first of all you expect the fire suppression system to solve the problem. certainly buy you more time but if you're on fire, the first

thing you do is radio for help and say, find me a runway to land on. so, even if the plane goes down in the ocean, that is -- if it goes down in the ocean under pilot control, that's a highly survivable event. the emergency slides double as rafts, and hasn't happened in a long time. people kind of forgot about it. but i don't think if you're on fire, you -- they call for help. so i don't -- and then you got imagine all those things i listed being knocked out and all the rather -- owl at once and all they're radio -- there's five radio ontario reason so you can't lose radio contact with them. >> there is way in the future they'll refigure out how that can't happen again, where you can't basically lose track of a plane for seven hours. >> no, because i think it's a criminal exact you -- just look the one piloted, german pilot

committed suicide. there's -- there nor easy solutions for that. don think you can -- the cockpits are locked in the united states because of fears for terrorism. i'm more afraid of a terrorism breaking in than of a crazy pilot. it's possible but, no, i don't think there's a solution to a criminal act. sorry. >> hi. it seems to me recently there's been a dramatic increase in military aircraft accidents including like few hours ago. a c-130 went down in georgia. is that just me or has there been that increase. >> those are off my radar screen because what i do is very much driven by ntsb, national transportation safety board, and basically a plane crashes anywhere in the world, they almost always put out a report in english. the world wants to know what

happened but not with military planes. so they have a different database if don't have any access to it. so they're playing in a different sand box. >> i had quick question. you mentioned briefly pilotless plains and that wouldn't be a good idea because of it's good to have a human pilot in the cockpit. can you talk more but the benefits and the pitfalls ol' the move towards, like, transportation that is not manned by people. >> well, i'm not a big fan of computers taking over the world. seems to be a generational thing. all my old innings engineering fends agree with me and the younger engineers act like it's already occurred. but what i did mention, there's -- well, my other

favorite example, imagine a truck wanders on the runway, the the absolute worst point of takeoff, already well established technology for a computer to measure the distance, and make a better, faster decision than the pilot, i don't think it's going to happen because it almost never happens so if you do the formal analysis that they do, they'll conclude you could justify spending essentially nothing and there's probably hundreds of things that almost never happen. it's easier to keep the pilot. so, i don't -- i won't say never. never is too long, but i don't know. communication link is no where near life and death reliable. who doesn't have problems with a cellphone? i don't think they're nowhere near life and death reliable. i steadied another failure, massive random damage. they waited 50 minutes to

complete all the check lifts with the help of the computer. the computer is the worriment don't want to write the computer out about the computer could not figure out how to land the plane. there's so much damage they couldn't calculate a confer racing that worked. the pilots had to cheat the computer to get it to approve a landism don't see how you write the humans out of the loop eve and how you due sign for massive random damage but that's just me. >> anymore questions? should we declare an end in thank you for coming. [applause] [inaudible conversations]

>> today between to bring you unfiltered coverage of congress the white house, the supreme court, and public policy events in washington, dc and around the country. c-span is brought to you by your cable or satellite provider. [inaudible conversations] good evening and welcome to community book store. i'm the co-owner and event